Hardeep Bansil (University of Birmingham) 2013, Antwerp 2-6 December 2013  Total Inelastic Cross Section [Nat. Commun. 2 (2011) 463, arXiv: ]Nat. Commun. 2 (2011) 463arXiv:  Rapidity Gap Cross Sections [Eur. Phys. J. C72 (2012) 1926, arXiv: ]Eur. Phys. J. C72 (2012) 1926arXiv:  Forward Transverse Energy Flow [JHEP11(2012)033, arXiv: ]JHEP11(2012)033arXiv: Overview of results on diffraction from the ATLAS Experiment

Single dissociation (SD), pp → Xp Double dissociation (DD), pp → XY Diffractive Dissociation ξ Y = M Y 2 /s At LHC energies, M X, M Y can range from m p +m π  ~1 TeV Diffractive channels together account for ~ quarter of total LHC cross section PYTHIA SD ξ (X) = M X 2 /s |t| < ~1 GeV N Ev [GeV -2 ] N Ev d|-t|

“Standard Model” of Soft Diffraction At fixed s: Deviations from this behaviour sensitive to α IP (t) and to absorptive corrections  c.f. multi-parton interactions i.e. approximately: X p Up to event-by-event hadronisation fluctuations,  variable predictable from empty rapidity regions … ~ flat gap distributions 3

ATLAS Acceptance Data obtained using full calorimeter coverage (|η| < 4.9) and inner tracking detector (|η| < 2.5) MBTS scintillators provide almost unbiased trigger Systems with M X (M Y ) < 7 GeV fall out of acceptance Detector is sensitive to particle production with p T > ~200 MeV 4

Total Inelastic pp Cross Section (ATLAS) 20 μb -1 sample selected using MBTS trigger (2.1 < |η| < 3.8), miss only elastic (pp→pp) and low mass diffraction (pp→pX etc) using limit of ξ > 5×10 -6 MBTS Unextrapolated result below PYTHIA and PHOJET defaults 5-15% extrapolation yields total inelastic cross section Extrapolation includes large uncertainty on low mass diffraction 5

Comparison with Subsequent Data Central Value of extrapolated ATLAS result consistent with TOTEM (due to large model dependence errors), but central value somewhat lower … need improved modelling of low mass diffraction... 6

Differential rapidity gap cross-sections Cross sections differential in “visible” rapidity gap size Δη F - Δη F extends from η= ±4.9 to first particle with p T > p T cut 200 MeV < p T cut < 800 MeV 0 < Δη F < 8 … corresponding (where diffraction dominates) to ~10 -6 < ξ < ~10 -2 … or ~7 < M x < ~700 GeV Corrected for experimental effects to level of stable hadrons p T cut = 200 MeV results follow … Δη F ~ 5.5 at p T cut = 200 MeV Implies ξ ~

Differential Rapidity Gap Cross Section Precision between ~8% (large gaps) and ~20% (Δη F ~ 1.5) Large gaps measure x-sec for SD [+ DD with M Y < ~7 GeV] Small gaps sensitive to hadronisation fluctuations / MPI … huge uncertainties PYTHIA best at small gaps, PHOJET > 50% high at Δη F ~ 1.5 8

9 Large Gaps and Diffractive Dynamics Diffractive plateau with ~1 mb per unit of gap size for Δη F > 3 broadly described by models PYTHIA high (DD much larger than in PHOJET) PHOJET low at high Δη F 9

Large Gaps and Diffractive Dynamics Default PHOJET and PYTHIA models have α IP (0) = 1 Donnachie-Landshoff flux has α IP (0) = Fit 6<Δη F < 8 region: α IP (0) = ± (stat) ± (sys) [Absorptive corrections neglected in all cases] ξ X ~ ξ X ~

Integrating ATLAS gap cross section up to some max Δη F (equivalently min ξ X ) and comparing with TOTEM indicates that small ξ X region underestimated in PHOJET and PYTHIA: 14 mb with ξ < 10 -5, compared to 6 (3) mb in PYTHIA (PHOJET) [Inelastic cross section excluding diffractive channels with ξ < ξ cut ] 11 Investigating Low Mass Extrapolations

12 Proton Spectrometry at ATLAS Rapidity gap measurements limited by several factors ALFA and AFP are the medium and long-term future of diffraction in ATLAS … will not be covered here … see talks in Friday’s session ALFA 12

Forward energy flow study of energy flow in soft and hard QCD processes MinBias sample (7.1 μb -1 ) – MBTS trigger, 1 vtx with at least 2 tracks > 250 MeV truth (> 150 MeV reco), no additional vertices Dijet sample (590 μb -1 ) - as above, study UE with back-to-back jets E T > 20 GeV, |η| 2.5, E T (J2)/E T (J1) > 0.5 Correct for experimental effects back to hadron level for particles with |η| 500 MeV, p neutral > 200 MeV) Transverse energy sum Average transverse energy density

Comparison in |η| bins 14 MinBias 0.0 < |η| < 0.8 UE Dijets MinBias 4.0 < |η| < 4.8 UE Dijets

Transverse Energy Density in |η| 15 MinBias UE Dijets - MinBias: Several models do acceptable job in central region - All models low for forward energy flow (emerging LHC theme) - Dedicated forward heavy ion / cosmic air shower model, EPOS, among best descriptions - UE dijets: 3 times higher energy flow than in minimum bias events - Similar conclusions, in particular poor forward description - Pythia 6 DW provides best overall description - [EPOS was never tuned on LHC UE data]

Transverse Energy Density Ratio 16 Dijet E T density larger than for minbias  hard scatter biases to small impact parameter. Sensitive to multi parton interactions. Reduction with |η| partly due to p cuts on particles in ΣE T (c.f no p cuts curve). Further decrease associated to additional UE in the central region from hard scatter.

Diffractive contribution 17 Compare nominal Pythia8 4C [N = 51; S = 12; D = 8] (mb) with [N; S/2; D/2] and [N; 2S; 2D]. Fewer particles in diffractive events: Changing contribution of diffractive components alters normalisation. Shape of E T density not significantly affected PDF contribution discussed in ATLAS small-x talk

Summary Precise soft diffractive and inelastic cross section data - Broadly described by single soft pomeron with intercept as expected - Low mass diffractive dissociation remains unresolved Forward energy flow generally underestimated by around 20-30%, with the exception of PYTHIA6 DW for dijet data and EPOS LHC for minbias data. - Diffractive processes reduce ‹E T density › but adjusting contributions leaves shapes largely unaffected … progress in first phase of LHC, but still a long way to go to completely understand diffractive physics. In particular, simultaneous description of diffractive / non-diffractive data in framework of multiple interactions / rapidity gap survival … 18

Back up slides 19

20 Uncertainties in pre-LHC Predictions Single dissociation σ = 14mb (PYTHIA8) σ = 10mb (PHOJET) Double dissociation σ = 9mb (PYTHIA8) σ = 4mb (PHOJET) sqrt(s) = 19.6 GeV SD DD [PHOJET: hep-ph/ ] Parameterisations based on old low energy data, particularly poor for DD 20

Increasing the p T cut defining gaps As p T cut increases, data shift to larger Δη F in a manner sensitive to hadronisation fluctuations and underlying event - Switching to p T cut = 400 MeV doesn’t change qualitative picture - Diffractive / non-diffractive processes barely distinguished at p T cut = 800 MeV 21

Small Gaps and Hadronisation - Big variation between MCs in small non-zero gap production via ND  fluctuations / UE - PYTHIA8 best at small gaps - PHOJET > 50% high at Δη F ~ See also higher p T cut data 22

Cluster Fragmentation: HERWIG++ - HERWIG++ with underlying event tune UE7-2 contains no explicit model of diffraction, but produces large gaps at higher than measured rate and a “bump” near Δη F = 6 - Effect not killed by removing colour reconnection or events with zero soft or semi-hard scatters in eikonal model Some investigations / progress since, but still not fully solved and remains a challenge 23

Durham Model of all Soft Diffractive Processes … simultaneous Durham (KMR) description of ATLAS gaps data and elastic cross section data from ISR to Totem based on a single pomeron in a 3-channel eikonal model, with significant absorptive corrections in gaps / dissociation case [arXiv: ] 24

Dijet production with veto on additional central jet activity Select events with two high p T jets, separated in rapidity by ∆y Veto on additional jet activity (with kT > Q 0, with Q 0 ≫ ΛQCD) between the two jets Q 0 taken as 20 GeV (not necessarily diffractive events) Measure “gap fraction”: dijet events with jet veto total dijet events Compared to HEJ at parton level and NLO POWHEG interfaced with PYTHIA and HERWIG 25

Dijet production with veto on additional central jet activity Gap fraction 26

Dijet production with veto on additional central jet activity Gap fraction 27